Abstract

Background: Married persons are healthier and live longer than single, divorced, and widowed persons. Time trends in self-rated health (SRH) by marital status and cohabitation have remained largely unstudied. We aim to assess the levels and trends of SRH by official marital status and cohabitation, and to study the causes of these differences. Methods: Two nationally representative cross-sectional surveys were conducted 20 years apart in Finland. Data on self-reported marital status, SRH, education, smoking, and long-standing illness were collected from Finns aged 30–64 years in 1978–80 (Mini-Finland Health Survey, N = 6102, response rate 96%) and 2000–01 (Health 2000 Survey, N = 5871, response rate 92%). Results: SRH has improved in the last 20 years, but differences between marital status groups have not reduced. In 2000–01, non-married persons reported worse SRH than married persons. Among men, single [cumulative odds ratio (COR) = 1.55; 95% confidence interval (95% CI) 1.22–1.99] and divorced (COR = 1.55; 95% CI 1.17–2.05) persons showed the poorest SRH, while among women widows (1.53; 95% CI 1.04–2.26) were the most disadvantaged group. The SRH of cohabiting persons did not significantly differ from that of married persons. Differences in educational structure, smoking, and the prevalence of long-term illness explain part of the marital status differences in SRH among men, but less so among women. Among both single men and women as well as among widowed women, SRH had improved slightly less than in the other groups. Conclusion: The challenges on public health posed by growing numbers of currently not married people are likely to increase.

William Farr studied the relationship between marital status and longevity in the 19th century,1 and the subject has intrigued researchers ever since. Mortality rates differ between marital status groups, the married being in the most advantaged position.25 Differences between marital status groups are also found in relation to several other health indicators, such as the prevalence of the most common diseases and various functional limitations.3,610 Non-married men and women consistently report poorer self-rated health (SRH) than married people do.912 Also, in health endangering behaviours married persons have the lowest rates while divorced persons tend to have the highest rates, and cohabiting, single, and widowed persons usually have prevalence in between.7,11,13

Two theories explain the relationship between marital status and health. According to the social causation theory, marriage is believed to have a health promoting effect whereas not being married has a negative effect on health. The suggested mechanisms include, for example, higher levels of material resources and social control of health-damaging behaviours such as smoking among married people.11,1416 On the other hand, the selection theory suggests that healthy people are much more likely to enter and maintain a relationship than those who are not healthy (direct selection) or that a variety of health-damaging behavioural and social factors increase the likelihood of both poor health and remaining or becoming unmarried (indirect selection).17,18 Although most researchers conclude that the effects work both ways,11,16 the relative importance of the different processes behind marital status disparities in health remains unclear.

Since the late 1970s, the relative mortality differences between married and non-married Finns have increased markedly, and by the turn of the century more than one-third of all deaths among middle-aged Finns were attributable to excess mortality among the non-married.19 The increase in the relative mortality differences is seen among men and women in all age groups between 30 and 89 years, and it is largely caused by a significant decline in mortality among married persons.3 Increasing health differences by marital status have also been observed in other countries.20 The contribution of the increased causal effects of marital status versus increased selection to the growing disparities in mortality is not known. There are very few analyses of time trends in morbidity differences between marital status groups in Finland or elsewhere.21 Furthermore, very little evidence exists on the levels and trends in the health of cohabiting men and women.

In this study, we examine differences in SRH between marital status and cohabiting groups in the middle-aged Finnish population and the time trends in these differences during the past 20 years. In addition, we assess the extent to which the level and change in education, smoking, and longstanding illness explain marital status differences in SRH and the changes in them.

Methods

Data

Our data were derived from two comparable cross-sectional studies conducted at a 20 year interval. The Mini-Finland Health Survey was carried out in 1978–80 and it was based on a nationally representative two stage cluster sample of Finns aged 30 years and over.22 Our study concerned subjects aged 30–64 years, and in this age-range the sample included 6102 subjects, of whom 5862 (96.1%) were interviewed. In 2000–01, another nationally representative extensive health survey, Health 2000, was carried out. In the Health 2000 survey a corresponding sample of participants aged 30–64 years was selected; altogether there were 5871 subjects, of whom 5405 (92.1%) were interviewed. The response rates for both surveys are presented in table 1. The methods of data collection in the Health 2000 survey were comparable to those in the Mini-Finland Health Survey.8

Table 1

Samples and participation rates by survey year, gender, and age-group

 1978–80 (Mini-Finland)
 
  2000–01 (Health 2000)
 
  

 
Total (%)
 
Men n (%)
 
Women n (%)
 
Total (%)
 
Men n (%)
 
Women n (%)
 
Sample 6102 (100) 2946 (100) 3156 (100) 5871 (100) 2891 (100) 2980 (100) 
Participationa 5862 (96.1) 2820 (95.7) 3042 (96.4) 5405 (92.1) 2621 (90.7) 2784 (93.4) 
Age-group       
    30–44 2795 (96.4) 1381 (95.4) 1414 (97.4) 2423 (90.8) 1172 (89.1) 1251 (92.5) 
    45–54 1666 (96.7) 814 (96.9) 852 (96.5) 1776 (92.5) 878 (91.2) 898 (93.8) 
    54–64 1401 (94.7) 625 (94.8) 776 (94.5) 1206 (94.1) 571 (93.3) 635 (94.8) 
 1978–80 (Mini-Finland)
 
  2000–01 (Health 2000)
 
  

 
Total (%)
 
Men n (%)
 
Women n (%)
 
Total (%)
 
Men n (%)
 
Women n (%)
 
Sample 6102 (100) 2946 (100) 3156 (100) 5871 (100) 2891 (100) 2980 (100) 
Participationa 5862 (96.1) 2820 (95.7) 3042 (96.4) 5405 (92.1) 2621 (90.7) 2784 (93.4) 
Age-group       
    30–44 2795 (96.4) 1381 (95.4) 1414 (97.4) 2423 (90.8) 1172 (89.1) 1251 (92.5) 
    45–54 1666 (96.7) 814 (96.9) 852 (96.5) 1776 (92.5) 878 (91.2) 898 (93.8) 
    54–64 1401 (94.7) 625 (94.8) 776 (94.5) 1206 (94.1) 571 (93.3) 635 (94.8) 

a: The numbers of participants exclude 47 persons with missing information on SRH (15) or on marital status (4), or ‘can not say’ (28) to SRH

Variables

SRH is a global measure of health, determined by health behaviour and illness.23,24 It also predicts future health problems25 and mortality.25,26 Furthermore, SRH correlates well with the subject's health status reported by a physician.22 Our outcome variable SRH was asked as follows: ‘How good is your health status today? Is it good, rather good, moderate, rather poor, or poor?’ In the Mini-Finland Health Study the measure of SRH had good reliability.27 Self-reported marital status was classified as married, cohabiting, single, divorced, or widowed.

Explanatory variables were chosen on the basis of earlier findings and hypotheses on the causes of marital status differences.9,11,16,23,28 Educational level indicates socioeconomic status, and it may affect both health and marital status through various mechanisms such as material resources and health behaviour. Smoking may both mediate the effects of marital status on health and act as a selective factor. Moreover, long-term illness is included in the analyses as it is one of the strongest determinants of SRH24 and it may contribute to marital status differences in SRH by acting as a selective or mediating factor.

Education was classified as basic (no matriculation examination and at most a vocational course or on the job training), intermediate, or higher education (university qualification). Smoking status was classified as never smoker; ex-regular-smoker; and current regular smoker of cigarettes, tobacco, or pipe. Current occasional smoking among ever-regular-smokers was asked only in the Health 2000 survey, and people in this category were classified as ex-regular-smokers. The question on longstanding illness was ‘Do you have any condition or longstanding illness that limits your work or functional ability?(Yes/No)’.

Statistical analyses

The analyses were carried out with the Survey procedures of Stata, which take the sampling design into account.29 We first calculated the distribution of the five-category SRH by survey year, gender, and marital status. Next, we analysed the association between the independent variables and the SRH by means of ordinal logistic regression.29,30 The results are presented as cumulative odds ratios (CORs) based on the odds of the poorer SRH [P(Y > j)/P(Y ≤ j) where j = 1, …, 4, 1 = good, …, 4 = rather poor, 5 = poor health]. The COR is the ratio of the odds of being in the poorer health category in a specific category of independent variables and the corresponding odds in the reference category. This model assumes that the COR of an independent variable does not depend on the threshold value of the SRH. We analysed the association between marital status and the five-category SRH separately for the two datasets. We sequentially adjusted for age, education, current smoking, and longstanding illness in order to assess their possible contribution to the differences between marital status groups. Finally, we combined the two datasets to test the statistical significance of changes in the marital status pattern of SRH (i.e. the interaction between marital status and period).

Results

The percentage of married and widowed persons declined substantially and a corresponding increase took place in the proportion of cohabiting, single, and divorced people among men and women during two decades (table 2). Among both genders, the percentage of those with high education grew. The proportion of men who had never smoked increased considerably and the percentage of current smokers decreased. In women, the trends in smoking were the opposite. Among both genders the proportion having at least one longstanding illness decreased.

Table 2

Age adjusteda distribution (%) of marital status, education, smoking, and longstanding illness among men and women aged 30–64 years in 1978–80 and 2000–01

 Men
 
 Women
 
 

 
1978–80
 
2000–01
 
1978–80
 
2000–01
 
Marital status     
    Married 79 59 74 60 
    Cohabiting 14 13 
    Single 12 16 11 
    Divorced 12 
    Widowed 
Education     
    Basic 67 31 69 28 
    Intermediate 26 58 25 58 
    High 11 13 
Smoking     
    Never smoked 29 38 73 59 
    Ex-smoker 32 30 11 19 
    Current smoker 40 32 16 21 
Longstanding illness     
    No 50 57 51 55 
    Yes 50 43 49 45 
N 2820 2621 3042 2784 
 Men
 
 Women
 
 

 
1978–80
 
2000–01
 
1978–80
 
2000–01
 
Marital status     
    Married 79 59 74 60 
    Cohabiting 14 13 
    Single 12 16 11 
    Divorced 12 
    Widowed 
Education     
    Basic 67 31 69 28 
    Intermediate 26 58 25 58 
    High 11 13 
Smoking     
    Never smoked 29 38 73 59 
    Ex-smoker 32 30 11 19 
    Current smoker 40 32 16 21 
Longstanding illness     
    No 50 57 51 55 
    Yes 50 43 49 45 
N 2820 2621 3042 2784 

a: Direct standardization, standard population = combined samples of both surveys, men and women together

In general, the percentage of good and rather good SRH increased, and correspondingly the rates of rather poor and poor health declined remarkably (table 3).

Table 3

Age adjusteda distribution (%) of SRH by marital status in 1978–80 and 2000–01, women and men aged 30–64 years

 Good health
 
 Rather good health
 
 Moderate health
 
 Rather poor health
 
 Poor health
 
 

 
1978–80
 
2000–01
 
1978–80
 
2000–01
 
1978–80
 
2000–01
 
1978–80
 
2000–01
 
1978–80
 
2000–01
 
Men           
    Marital status           
        Married 26 40 26 30 34 23 
        Cohabiting 19 41 30 30 35 19 13 
        Single 25 31 25 29 33 29 13 
        Divorced 23 33 22 28 27 24 16 12 
        Widowedb 14 16 21 63 39 13 13 13 
        All 25 38 26 30 34 24 10 
Women           
    Marital status           
        Married 28 42 29 32 32 21 10 
        Cohabiting 13 37 34 33 31 24 16 
        Single 30 40 32 29 26 23 10 
        Divorced 30 38 23 30 32 22 11 
        Widowed 27 36 31 20 31 36 
        All 28 40 29 32 31 22 
 Good health
 
 Rather good health
 
 Moderate health
 
 Rather poor health
 
 Poor health
 
 

 
1978–80
 
2000–01
 
1978–80
 
2000–01
 
1978–80
 
2000–01
 
1978–80
 
2000–01
 
1978–80
 
2000–01
 
Men           
    Marital status           
        Married 26 40 26 30 34 23 
        Cohabiting 19 41 30 30 35 19 13 
        Single 25 31 25 29 33 29 13 
        Divorced 23 33 22 28 27 24 16 12 
        Widowedb 14 16 21 63 39 13 13 13 
        All 25 38 26 30 34 24 10 
Women           
    Marital status           
        Married 28 42 29 32 32 21 10 
        Cohabiting 13 37 34 33 31 24 16 
        Single 30 40 32 29 26 23 10 
        Divorced 30 38 23 30 32 22 11 
        Widowed 27 36 31 20 31 36 
        All 28 40 29 32 31 22 

a: Direct standardization, standard population = combined samples of both surveys, men and women together

b: Age categories (30–34) and (35–39) were combined when adjusting for age

As a whole, the age-adjusted differences in SRH between men's marital status groups in 1978–80 did not quite achieve statistical significance (table 4). However, divorced men reported the worst SRH (COR = 1.74; 1.18–2.55) compared with married men, and also single and cohabiting men showed a mild excess morbidity. The excess morbidity of divorced men was partly explained by their low level of education. In 2000–01, the difference between marital status groups among men was highly significant. The SRH of cohabiting men or widowers did not significantly differ from that of married persons. Single (COR = 1.55; 1.22–1.99) and divorced (COR = 1.55; 1.17–2.05) men had the worst health. The poor SRH of single men in 2000–01 is mostly explained by their low education and high prevalence of longstanding illness, whereas education and smoking explained a considerable proportion of the poor SRH of divorced men.

Table 4

Cumulative odds ratios (CORs) for 5-class SRH (1 = good health, …, 5 = poor health) by marital status, education, smoking, and longstanding illness, men aged 30–64 years (with 95% CI)


 
Agea
 
+Education
 
+Smoking
 
+Illness
 
1978–80z (Mini-Finland)     
    Marital status     
        Married 1.00 1.00 1.00 1.00 
        Cohabiting 1.16 (0.81–1.66) 1.13 (0.80–1.59) 1.08 (0.73–1.53) 0.93 (0.62–1.41) 
        Single 1.20 (0.95–1.51) 1.08 (0.87–1.35) 1.06 (0.86–1.32) 0.94 (0.74–1.19) 
        Divorced 1.74 (1.18–2.55) 1.56 (1.05–2.31) 1.51 (1.02–2.24) 1.51 (1.02–2.4) 
        Widowed 0.92 (0.45–1.89) 0.82 (0.39–1.72) 0.81 (0.38–1.72) 0.65 (0.28–1.51) 
    Pb ns ns ns ns 
    Education     
        High 1.00 1.00 1.00 1.00 
        Intermediate 2.48 (1.82–3.37) 2.46 (1.77–3.42) 2.29 (1.64–3.21) 2.06 (1.46–2.91) 
        Basic 3.80 (2.78–5.20) 3.73 (2.65–5.25) 3.45 (2.44–4.89) 2.95 (2.09–4.16) 
    Pb <0.001 <0.001 <0.001 <0.001 
    Smoking     
        Never smoked 1.00  1.00 1.00 
        Ex-smoker 1.35 (1.20–1.52)  1.29 (1.10–1.51) 1.22 (1.02–1.46) 
        Current smoker 1.56 (1.33–1.83)  1.41 (1.18–1.67) 1.45 (1.19–1.75) 
    Pb <0.001  <0.001 <0.001 
    Longstanding illness     
        No 1.00   1.00 
        Yes 7.21 (6.38–8.16)   7.12 (6.17–8.22) 
        Pb <0.001   <0.001 
2000–01 (Health 2000)     
    Marital status     
        Married 1.00 1.00 1.00 1.00 
        Cohabiting 1.08 (0.86–1.37) 1.00 (0.79–1.26) 0.97 (0.77–1.23) 0.95 (0.75–1.21) 
        Single 1.55 (1.22–1.99) 1.36 (1.05–1.75) 1.31 (1.01–1.69) 1.16 (0.88–1.54) 
        Divorced 1.55 (1.17–2.05) 1.42 (1.07–1.89) 1.30 (0.98–1.73) 1.26 (0.93–1.70) 
        Widowed 1.04 (0.53–2.06) 0.97 (0.50–1.89) 0.99 (0.51–1.93) 0.82 (0.39–1.72) 
    Pb 0.001 <0.05 ns ns 
    Education     
        High 1.00 1.00 1.00 1.00 
        Intermediate 2.10 (1.64–2.68) 2.04 (1.59–2.60) 1.91 (1.49–2.44) 1.64 (1.27–2.12) 
        Basic 3.45 (2.63–4.52) 3.25 (2.47–4.27) 2.92 (2.23–3.82) 2.37 (1.79–3.14) 
    Pb <0.001 <0.001 <0.001 <0.001 
    Smoking     
        Never smoked 1.00  1.00 1.00 
        Ex-smoker 1.12 (0.95–1.32)  1.04 (0.88–1.22) 1.04 (0.87–1.25) 
        Current smoker 1.87 (1.53–2.28)  1.58 (1.30–1.93) 1.64 (1.34–2.01) 
    Pb <0.001  <0.001 <0.001 
    Longstanding illness     
        No 1.00   1.00 
        Yes 5.22 (4.43–6.15)   4.99 (4.23–5.89) 
    Pb <0.001   <0.001 

 
Agea
 
+Education
 
+Smoking
 
+Illness
 
1978–80z (Mini-Finland)     
    Marital status     
        Married 1.00 1.00 1.00 1.00 
        Cohabiting 1.16 (0.81–1.66) 1.13 (0.80–1.59) 1.08 (0.73–1.53) 0.93 (0.62–1.41) 
        Single 1.20 (0.95–1.51) 1.08 (0.87–1.35) 1.06 (0.86–1.32) 0.94 (0.74–1.19) 
        Divorced 1.74 (1.18–2.55) 1.56 (1.05–2.31) 1.51 (1.02–2.24) 1.51 (1.02–2.4) 
        Widowed 0.92 (0.45–1.89) 0.82 (0.39–1.72) 0.81 (0.38–1.72) 0.65 (0.28–1.51) 
    Pb ns ns ns ns 
    Education     
        High 1.00 1.00 1.00 1.00 
        Intermediate 2.48 (1.82–3.37) 2.46 (1.77–3.42) 2.29 (1.64–3.21) 2.06 (1.46–2.91) 
        Basic 3.80 (2.78–5.20) 3.73 (2.65–5.25) 3.45 (2.44–4.89) 2.95 (2.09–4.16) 
    Pb <0.001 <0.001 <0.001 <0.001 
    Smoking     
        Never smoked 1.00  1.00 1.00 
        Ex-smoker 1.35 (1.20–1.52)  1.29 (1.10–1.51) 1.22 (1.02–1.46) 
        Current smoker 1.56 (1.33–1.83)  1.41 (1.18–1.67) 1.45 (1.19–1.75) 
    Pb <0.001  <0.001 <0.001 
    Longstanding illness     
        No 1.00   1.00 
        Yes 7.21 (6.38–8.16)   7.12 (6.17–8.22) 
        Pb <0.001   <0.001 
2000–01 (Health 2000)     
    Marital status     
        Married 1.00 1.00 1.00 1.00 
        Cohabiting 1.08 (0.86–1.37) 1.00 (0.79–1.26) 0.97 (0.77–1.23) 0.95 (0.75–1.21) 
        Single 1.55 (1.22–1.99) 1.36 (1.05–1.75) 1.31 (1.01–1.69) 1.16 (0.88–1.54) 
        Divorced 1.55 (1.17–2.05) 1.42 (1.07–1.89) 1.30 (0.98–1.73) 1.26 (0.93–1.70) 
        Widowed 1.04 (0.53–2.06) 0.97 (0.50–1.89) 0.99 (0.51–1.93) 0.82 (0.39–1.72) 
    Pb 0.001 <0.05 ns ns 
    Education     
        High 1.00 1.00 1.00 1.00 
        Intermediate 2.10 (1.64–2.68) 2.04 (1.59–2.60) 1.91 (1.49–2.44) 1.64 (1.27–2.12) 
        Basic 3.45 (2.63–4.52) 3.25 (2.47–4.27) 2.92 (2.23–3.82) 2.37 (1.79–3.14) 
    Pb <0.001 <0.001 <0.001 <0.001 
    Smoking     
        Never smoked 1.00  1.00 1.00 
        Ex-smoker 1.12 (0.95–1.32)  1.04 (0.88–1.22) 1.04 (0.87–1.25) 
        Current smoker 1.87 (1.53–2.28)  1.58 (1.30–1.93) 1.64 (1.34–2.01) 
    Pb <0.001  <0.001 <0.001 
    Longstanding illness     
        No 1.00   1.00 
        Yes 5.22 (4.43–6.15)   4.99 (4.23–5.89) 
    Pb <0.001   <0.001 

a: Model: SRH = age + variable (marital status; education; smoking; longstanding illness)

b: The statistical significance of the difference of the CORs between the categories of the variable in question based on the Wald test

Among women there were significant differences in SRH between the marital status groups in 1978–80, resulting largely from the poor SRH of cohabiting women (COR = 1.85; 1.29–2.63) (table 5). The SRH of single and widowed women was even slightly better than that of married women (no statistical significance). The excess morbidity of cohabiting women remained statistically significant after adjusting for educational structure, smoking, and long-term illness although these factors reduced the excess by almost one-third. In 2000–01 the SRH differences by marital status did not quite reach statistical significance and their pattern was different from that in 1978–80. In 2000–01, married women had better SRH compared with all non-married groups, and the widowed women had the worst SRH (COR = 1.53; 1.04–2.26). The excess morbidity of the widowed women was largely explained by their longstanding illnesses and educational structure, whereas controlling for these two factors and smoking rather accentuated than reduced the slight excess morbidity of single, cohabiting, and divorced women compared with married women in 2000–01.

Table 5

CORs for 5-class SRH (1 = good health, …, 5 = poor health) by marital status, education, smoking, and longstanding illness, women aged 30–64 years (with 95% CI)


 
Agea
 
+Education
 
+Smoking
 
+Illness
 
1978–80 (Mini-Finland)     
    Marital status     
        Married 1.00 1.00 1.00 1.00 
        Cohabiting 1.85 (1.29–2.63) 1.75 (1.21–2.54) 1.68 (1.14–2.47) 1.61 (1.02–2.53) 
        Single 0.84 (0.66–1.06) 0.90 (0.70–1.14) 0.88 (0.69–1.12) 0.79 (0.61–1.04) 
        Divorced 1.13 (0.84–1.53) 1.12 (0.84–1.50) 1.08 (0.80–1.46) 1.01 (0.74–1.38) 
        Widowed 0.84 (0.61–1.16) 0.82 (0.59–1.14) 0.80 (0.57–1.12) 0.86 (0.62–1.19) 
    Pb <0.01 <0.05 ns ns 
    Education     
        High 1.00 1.00 1.00 1.00 
        Intermediate 1.92 (1.41–2.60) 1.88 (1.38–2.57) 1.89 (1.38–2.59) 1.59 (1.12–2.25) 
        Basic 3.41 (2.61–4.46) 3.34 (2.37–4.71) 3.33 (2.35–4.70) 2.80 (2.00–3.99) 
    Pb <0.001 <0.001 <0.001 <0.001 
    Smoking     
        Never smoked 1.00  1.00 1.00 
        Ex-smoker 0.96 (0.76–1.22)  0.96 (0.74–1.23) 0.97 (0.76–1.24) 
        Current smoker 1.32 (1.16–1.55)  1.21 (1.01–1.45) 1.23 (1.00–1.51) 
    Pb <0.01  ns ns 
    Longstanding illness     
        No 1.00   1.00 
        Yes 7.31 (5.87–9.11)   7.15 (6.01–8.51) 
    Pb <0.001   <0.001 
2000–01 (Health 2000)     
    Marital status     
        Married 1.00 1.00 1.00 1.00 
        Cohabiting 1.22 (0.99–1.50) 1.18 (0.96–1.46) 1.14 (0.92–1.41) 1.26 (0.99–1.59) 
        Single 1.19 (0.90–1.57) 1.24 (0.94–1.65) 1.21 (0.91–1.60) 1.28 (0.95–1.72) 
        Divorced 1.20 (0.94–1.53) 1.19 (0.94–1.50) 1.13 (0.89–1.44) 1.24 (0.97–1.57) 
        Widowed 1.53 (1.04–2.26) 1.42 (0.96–2.11) 1.40 (0.94–2.09) 1.17 (0.78–1.74) 
    Pb ns ns ns ns 
    Education     
        High 1.00 1.00 1.00 1.00 
        Intermediate 1.85 (1.48–2.31) 1.85 (1.48–2.31) 1.79 (1.43–2.24) 1.65 (1.31–2.08) 
        Basic 2.71 (2.11–3.48) 2.69 (2.09–3.48) 2.54 (1.95–3.31) 2.18 (1.66–2.87) 
    Pb <0.001 <0.001 <0.001 <0.001 
    Smoking     
        Never smoked 1.00  1.00 1.00 
        Ex-smoker 0.96 (0.80–1.15)  0.93 (0.78–1.12) 0.92 (0.77–1.10) 
        Current smoker 1.52 (1.26–1.83)  1.31 (1.07–1.61) 1.34 (1.08–1.66) 
    Pb <0.001  <0.05 <0.01 
    Longstanding illness     
        No 1.00   1.00 
        Yes 4.88 (4.14–5.75)   4.81 (4.07–5.69) 
    P <0.001   <0.001 

 
Agea
 
+Education
 
+Smoking
 
+Illness
 
1978–80 (Mini-Finland)     
    Marital status     
        Married 1.00 1.00 1.00 1.00 
        Cohabiting 1.85 (1.29–2.63) 1.75 (1.21–2.54) 1.68 (1.14–2.47) 1.61 (1.02–2.53) 
        Single 0.84 (0.66–1.06) 0.90 (0.70–1.14) 0.88 (0.69–1.12) 0.79 (0.61–1.04) 
        Divorced 1.13 (0.84–1.53) 1.12 (0.84–1.50) 1.08 (0.80–1.46) 1.01 (0.74–1.38) 
        Widowed 0.84 (0.61–1.16) 0.82 (0.59–1.14) 0.80 (0.57–1.12) 0.86 (0.62–1.19) 
    Pb <0.01 <0.05 ns ns 
    Education     
        High 1.00 1.00 1.00 1.00 
        Intermediate 1.92 (1.41–2.60) 1.88 (1.38–2.57) 1.89 (1.38–2.59) 1.59 (1.12–2.25) 
        Basic 3.41 (2.61–4.46) 3.34 (2.37–4.71) 3.33 (2.35–4.70) 2.80 (2.00–3.99) 
    Pb <0.001 <0.001 <0.001 <0.001 
    Smoking     
        Never smoked 1.00  1.00 1.00 
        Ex-smoker 0.96 (0.76–1.22)  0.96 (0.74–1.23) 0.97 (0.76–1.24) 
        Current smoker 1.32 (1.16–1.55)  1.21 (1.01–1.45) 1.23 (1.00–1.51) 
    Pb <0.01  ns ns 
    Longstanding illness     
        No 1.00   1.00 
        Yes 7.31 (5.87–9.11)   7.15 (6.01–8.51) 
    Pb <0.001   <0.001 
2000–01 (Health 2000)     
    Marital status     
        Married 1.00 1.00 1.00 1.00 
        Cohabiting 1.22 (0.99–1.50) 1.18 (0.96–1.46) 1.14 (0.92–1.41) 1.26 (0.99–1.59) 
        Single 1.19 (0.90–1.57) 1.24 (0.94–1.65) 1.21 (0.91–1.60) 1.28 (0.95–1.72) 
        Divorced 1.20 (0.94–1.53) 1.19 (0.94–1.50) 1.13 (0.89–1.44) 1.24 (0.97–1.57) 
        Widowed 1.53 (1.04–2.26) 1.42 (0.96–2.11) 1.40 (0.94–2.09) 1.17 (0.78–1.74) 
    Pb ns ns ns ns 
    Education     
        High 1.00 1.00 1.00 1.00 
        Intermediate 1.85 (1.48–2.31) 1.85 (1.48–2.31) 1.79 (1.43–2.24) 1.65 (1.31–2.08) 
        Basic 2.71 (2.11–3.48) 2.69 (2.09–3.48) 2.54 (1.95–3.31) 2.18 (1.66–2.87) 
    Pb <0.001 <0.001 <0.001 <0.001 
    Smoking     
        Never smoked 1.00  1.00 1.00 
        Ex-smoker 0.96 (0.80–1.15)  0.93 (0.78–1.12) 0.92 (0.77–1.10) 
        Current smoker 1.52 (1.26–1.83)  1.31 (1.07–1.61) 1.34 (1.08–1.66) 
    Pb <0.001  <0.05 <0.01 
    Longstanding illness     
        No 1.00   1.00 
        Yes 4.88 (4.14–5.75)   4.81 (4.07–5.69) 
    P <0.001   <0.001 

a: Model: SRH = age + variable (marital status; education; smoking; longstanding illness)

b: The statistical significance of the difference of the CORs between the categories of the variable in question based on the Wald test

In general, poor SRH was clearly less common in 2000–01 than 20 years earlier; the age-adjusted COR describing the relative odds for poor health in 2000–01 in comparison with 1978–80 was 0.52 [95% confidence interval (95% CI) 0.46–0.58] among men and 0.50 (0.44–0.55) among women. According to the test for the interaction between marital status and period, the changes in the marital status patterns of SRH between the two periods—or, equivalently, the differences in the relative rate of change in SRH—were not statistically significant for either gender (P = 0.231 for men and P = 0.097 for women). However, in accordance with the findings in tables 4 and 5, the results suggest that the improvement in SRH has not been quite as substantial among single men (COR = 0.67; 95% CI 0.48–0.87) and single (COR = 0.66; 95% CI 0.43–0.89) or widowed women (COR = 0.64; 95% CI 0.34–0.94) as in the other marital status groups, while among cohabiting women SRH has improved slightly faster (COR = 0.35; 95% CI 0.21–0.48) than in other female groups.

Discussion

Summary of the findings

In accordance with previous studies, we found that in 2000–01 all non-married groups reported worse SRH than the married groups.9,11 Among men, single and divorced persons showed the poorest SRH, while among women widows were the most disadvantaged group. Low educational level and a high prevalence of longstanding illness explain a major part of the excess morbidity for single men and widowed women in comparison with married persons, while differences in the prevalence of smoking, together with those in educational structure, made a considerable contribution to the excess morbidity of divorced men. The SRH of cohabiting persons did not significantly differ from that of married persons.

In Finland SRH has improved in all marital status groups, but among single men and women and among widowed women SRH improved slightly less than in the other groups. Moreover, among cohabiting women the rate of improvement was higher than in other groups. As a result, the pattern of health differences among women has changed towards that of men.

Methodological considerations

The response rates in the two surveys were exceptionally high and the study protocols were very similar.8 In both studies, marital status was based on self-report, which made it possible to separate cohabiters from other groups. In a study on cohabitation in Finland in the 1970s, the proportion of cohabiting people was 6% in the age-group 25–44 years and 2% in the age-group 45–64 years.31 On this basis, we would have expected the proportion of cohabiters to be ∼4% in our data for the period 1978–80. However, we found a lower prevalence of cohabiters, particularly among women (2%), and also among men (3%). There may thus be underreporting of cohabitation in the Mini-Finland survey (1978–80), arising perhaps from the fact that spontaneously reported marital status may favour one's official marital status rather than the less common and at that time to some extent socially stigmatizing cohabitating. In the Health 2000 survey the marital status categories were read out loud before answering. This is expected to reduce such a bias, together with the more widespread social acceptability of cohabitation in current Finnish society. Accordingly, the prevalence of cohabiting obtained from the Health 2000 survey corresponds well with estimates by Statistics Finland (14% for men and 13% for women in our data versus 13 and 12%, respectively, in official family statistics for 1999).32 It is unclear whether the underestimation of the prevalence of cohabiting at the 1978–80 survey has biased our estimates of change in SRH among cohabiters.

Differences in SRH between marital status groups in 2000–01

Our findings support earlier results reporting poor SRH of non-married subjects.9,11 Marital status may directly affect health by means of economic support, psychosocial factors, and health behaviours. Our study shows that educational structure had an important role in explaining the current poor SRH of single or divorced men as expected on the basis of the lower average education of non-married men.3,28

Marriage has also been found to protect individuals from health-damaging behaviours such as smoking and alcohol consumption, and to promote healthy habits such as physical activity.9,11,33 The healthy lifestyle of married individuals may be partly connected with current or anticipated parenthood. Future fathers report a reduction of alcohol consumption and smoking at the onset of their wives' pregnancies.34 Joung et al.11 controlled for smoking, alcohol consumption, and leisure exercise, and reported this to attenuate the excess poor SRH among non-married groups. Our results support the hypothesis of the role of health behaviour in shaping health differences by marital status as smoking contributed to the current excess morbidity of divorced men. Smoking has been shown to strongly associate with other unhealthy behaviours.35

We found that adjusting for longstanding illness attenuated the excess poor SRH of single men and widowed women as compared with the married group. The group of widowed middle-aged women is currently very small, and both causation and selection may operate here. On one hand, bereavement at this age is a rare and stressful event, which may contribute to mental and somatic health.36,37 On the other hand, premature death of the spouse may reflect health and social problems in the family and health problems may have arisen before widowhood. Similar processes may be operating in the case of single males as well. As we had no information on the time of onset of illness, we cannot make causal inferences on the contribution of longstanding illness in either direction.

We found that currently the SRH of male cohabiters is practically the same as that of married men and better than that of single or divorced men. This was roughly the situation 20 years ago among men while cohabiting women used to show higher rates of poor SRH compared with married women. In our data, part of this difference was explained by the less favourable educational structure and a higher prevalence of smoking and longstanding illness among cohabiting women. These results are in line with previous studies reporting more health-damaging behaviours among cohabiters than the married group.11,38

Overall, education, smoking, and longstanding illness were important in explaining the health differences among men. However, the current slight excess morbidity of cohabiting, single, or divorced women in comparison with married women only accentuated after controlling for these three potential explanatory factors. This suggests that other causal processes are operating among women.

Time trends in SRH by marital status

In Finland the relative excess mortality has increased among non-married groups, and this is only partly explained by their becoming more marginalized in terms of socioeconomic position.3 Increasing relative mortality differences have also been documented in other Western countries.20,39

To our knowledge, there are no previous studies on changes in SRH by marital status directly comparable with our study. Elstad's findings on 31- to 60-year-old Norwegians suggest that women's health differences by marital status may have diminished between the early 1970s and the late 1980s.21 However, the time periods of study, adjusted variables, and the health measure are not directly comparable with ours. According to the results from annual surveys on health behaviour in the Finnish adult population since 1978, at the end of the 1970s divorced men had the worst SRH,40 whereas in 2000 single men had the worst SRH.41 Women's SRH did not significantly vary by marital status in either time period. Among men, the results are in line with ours. For women, however, our results suggest greater marital status differences in SRH especially in the 1978–80 period, when cohabiting women had the worst SRH according to our study. The lower response rate in the annual health behaviour surveys, ∼80% compared with ∼95% in our data, makes us assume that our results reflect the reality more accurately.

During the past 20 years, the relative position of cohabiting women in terms of SRH has clearly improved. As mentioned above, there may be underreporting of cohabiters in our data representing the late 1970s, but we find it unlikely that healthy persons would have been less likely to self-report cohabitation. Cohabiters may experience less social pressure than 20 years ago, which might contribute to the particularly rapid improvement in their SRH. In Finland, the proportion of cohabiters having children has doubled between the years 1979 and 1999 (age-group 15–64),32,42 which supports our suggestion that cohabitation today resembles marriage more than it did 20 years ago.

While cohabitation is much more common today than it was 20 years ago, the reverse is true for widowhood at middle-age. The worsening of the relative position of widows suggested by our results may thus partly result from the accentuation of both causative and selective processes discussed above in the case of cross-sectional differences. The improvement in the SRH of the single men and women has been slower than average, despite the growing relative size of this group, possibly indicating an increase in the effects of causal processes.

Conclusion

SRH in Finland has improved in the last 20 years, but the rate of improvement has been slightly slower among single men and women and widowed women, whereas the relative position of cohabiting women has improved. The marital status pattern of health differences among women has changed towards that of men, with the married showing the best SRH. Differences in educational structure, smoking, and the prevalence of long-term illness explain part of the current differences among men, but less so among women. Challenges to public health posed by growing numbers of currently not married persons are likely to increase. Understanding the relationship between SRH and marital status provides a basis for effective preventive health interventions aimed at the most vulnerable marital status groups.

Key points

  • Previous studies show marital status to be strongly associated with health, but the effects of cohabitation and time trends of these differences are poorly known.

  • In Finland self-rated health has improved in the last 20 years, but the excess morbidity of non-married groups has not reduced.

  • In 2000–01 single and divorced men and widowed women showed poorest self-rated health while SRH of cohabiters did not differ from that of the married group.

  • Education, smoking, and long-term illnesses explained part of the differences in SRH among men but less so among women.

  • The challenges on public health posed by growing numbers of currently not married people are likely to increase.

The authors thank the personnel of the Health 2000 Survey and the Mini-Finland Health Examination Survey. This work was funded by a grant from the Academy of Finland (No. 203418). The study is supported by a fellowship to Pekka Martikainen (No. 70631).

References

1
Farr W. Transactions of the National Association for the Promotion of Social Science. London: Parker,
1858
.
2
Cheung YB. Marital status and mortality in British women: a longitudinal study.
Int J Epidemiol
 
2000
;
29
:
93
–9.
3
Martikainen P, Martelin M, Koskinen S, Nihtilä E, Majamaa K. Differences in mortality by marital status in Finland from 1976 to 2000: Analyses of changes in marital status distributions, socio-demographic and household composition, and cause of death.
Popul Stud
 
2005
;
59
:
99
–115.
4
Lund R, Holstein BE, Osler M. Marital history from age 15 to 40 years and subsequent 10-year mortality: a longitudinal study of Danish males born in 1953.
Int J Epidemiol
 
2004
;
33
:
389
–97.
5
Hu YR, Goldman N. Mortality differentials by marital status: an international comparison.
Demography
 
1990
;
27
:
233
–50.
6
Murphy M, Glaser K, Grundy E. Marital status and long-term illness in Great Britain.
J Marriage Fam
 
1997
;
59
:
156
–64.
7
Mendes de Leon CF, Appels AW, Otten FW, Schouten EG. Risk of mortality and coronary heart disease by marital status in middle-aged men in The Netherlands.
Int J Epidemiol
 
1992
;
21
:
460
–6.
8
Aromaa A, Koskinen S, editors. Health and functional capacity in Finland. Baseline results of the Health 2000 health examination survey. Publications of the National Public Health Institute B12/2004. Helsinki: Hakapaino Oy,
2004
.
9
Helmert U, Shea S. Family status and self-reported health in West Germany.
Soz Praventivmed
 
1998
;
43
:
124
–32.
10
Joung IM, van de Mheen H, Stronks K, van Poppel FW, Mackenbach JP. Differences in self-reported morbidity by marital status and by living arrangement.
Int J Epidemiol
 
1994
;
23
:
91
–7.
11
Joung IM, Stronks K, van de Mheen H, Mackenbach JP. Health behaviours explain part of the differences in self reported health associated with partner/marital status in The Netherlands.
J Epidemiol Community Health
 
1995
;
49
:
482
–8.
12
Ren XS. Marital status and quality of relationships: the impact on health perception.
Soc Sci Med
 
1997
;
44
:
241
–9.
13
Ebrahim S, Wannamethee G, McCallum A, Walker M, Shaper AG. Marital status, change in marital status, and mortality in middle-aged British men.
Am J Epidemiol
 
1995
;
142
:
834
–42.
14
Kumlin L, Latscha G, Orth-Gomer K, et al. Marital status and cardiovascular risk in French and Swedish automotive industry workers—cross sectional results from the Renault-Volvo Coeur study.
J Intern Med
 
2001
;
249
:
315
–23.
15
Hughes ME, Waite LJ. Health in household context: living arrangements and health in late middle age.
J Health Soc Behav
 
2002
;
43
:
1
–21.
16
Wyke S, Ford G. Competing explanations for associations between marital status and health.
Soc Sci Med
 
1992
;
34
:
523
–32.
17
Mastekaasa A. Marriage and psychological well-being: some evidence on selection into marriage.
J Marriage Fam
 
1992
;
54
:
901
–11.
18
Joung IM, van de Mheen HD, Stronks K, van Poppel FW, Mackenbach JP. A longitudinal study of health selection in marital transitions.
Soc Sci Med
 
1998
;
46
:
425
–35.
19
Koskinen S, Martelin T, Rissanen H. Mortality differences by marital status—a growing public health problem (in Finnish). Sosiaalilääketieteellinen aikakausilehti.
J Soc Med
 
1999
;
36
:
271
–84.
20
Valkonen T. Trends in marital status differences in mortality. In: Vallin J, Mesle F, Valkonen T, editors. Trends in mortality and differential mortality. Strasbourg: Council of Europe Publishing,
2004
: 252–78.
21
Elstad JI. Inequalities in health related to women's marital, parental, and employment status—a comparison between the early 70s and the late 80s, Norway.
Soc Sci Med
 
1996
;
42
:
75
–89.
22
Aromaa A, Heliövaara M, Impivaara O, et al. Health, functional limitations and the need for care in Finland. Basic results from the Mini-Finland Health Survey. Helsinki and Turku: Publications of the Social Insurance Institution,
1989
.
23
Manderbacka K, Lundberg O, Martikainen P. Do risk factors and health behaviours contribute to self-ratings of health?
Soc Sci Med
 
1999
;
48
:
1713
–20.
24
Goldman N, Glei D, Chang M-C. The role of clinical risk factors in understanding self-rated health.
Ann Epidemiol
 
2003
;
14
:
49
–57.
25
Idler EL, Russell LB, Davis D. Survival, functional limitations, and self-rated health in the NHANES I Epidemiologic Follow-up Study, 1992. First National Health and Nutrition Examination Survey.
Am J Epidemiol
 
2000
;
152
:
874
–83.
26
Bardage C, Isacson D, Pedersen NL. Self-rated health as a predictor of mortality among persons with cardiovascular disease in Sweden.
Scand J Public Health
 
2001
;
29
:
13
–22.
27
Martikainen P, Aromaa A, Heliovaara M, et al. Reliability of perceived health by sex and age.
Soc Sci Med
 
1999
;
48
:
1117
–22.
28
Koskinen S, Martelin T. Why are socioeconomic mortality differences smaller among women than among men?
Soc Sci Med
 
1994
;
38
:
1385
–96.
29
StataCorp. Stata Statistical Software: Release 8.0. College Station, Texas: Stata Corporation,
2003
.
30
Rothman K, Greenland S. Modern epidemiology. Philadelphia: Lippincott-Raven Publishers,
1998
.
31
Aromaa K, Cantell I, Jaakkola R. Cohabitation in Finland in the 1970s. Valtion painatuskeskus, Helsinki: Research Institute of Legal Policy,
1983
.
32
Families 1999 (in Finnish). Helsinki: Statistics Finland,
2000
.
33
Miller-Tutzauer C, Leonard KE, Windle M. Marriage and alcohol use: a longitudinal study of “maturing out”.
J Stud Alcohol
 
1991
;
52
:
434
–40.
34
Hyssala L, Rautava P, Sillanpaa M, Tuominen J. Changes in the smoking and drinking habits of future fathers from the onset of their wives' pregnancies.
J Adv Nurs
 
1992
;
17
:
849
–54.
35
Laaksonen M, Prattala R, Karisto A. Patterns of unhealthy behaviour in Finland.
Eur J Public Health
 
2001
;
11
:
294
–300.
36
Martikainen P, Valkonen T. Mortality after the death of a spouse: rates and causes of death in a large Finnish cohort.
Am J Public Health.
 
1996
;
86
:
1087
–93.
37
Wilcox S, Evenson KR, Aragaki A, Wassertheil-Smoller S, Mouton CP, Loevinger BL. The effects of widowhood on physical and mental health, health behaviors, and health outcomes: The Women's Health Initiative.
Health Psychol
 
2003
;
22
:
513
–22.
38
Power C, Rodgers B, Hope S. Heavy alcohol consumption and marital status: disentangling the relationship in a national study of young adults.
Addiction
 
1999
;
94
:
1477
–87.
39
Valkonen T, Martikainen P, Blomgren J. Increasing excess mortality among nonmarried elderly people in developed countries.
Demogr Res
 
2004
. Special Collection 2: Article 12:
305
–30.
40
Koskinen S, Puska P, Valkonen T. Terveyskäyttäytyminen keski-ikäisen väestön osaryhmissä 1978–80 (in Finnish). Helsinki: Valtion painatuskeskus,
1981
.
41
Helakorpi S, Uutela A, Prättälä R, Puska P. Health behaviour and health among Finnish adult population, Spring 2000. Helsinki: Publications of the National Public Health Institute,
2000
.
42
Labour force survey 1978 (in Finnish). Helsinki: Statistics Finland,
1979
.

Comments

0 Comments